They fell in value (below $1 billion, in all) because deals like Intel’s $15.3 billion first-quarter offer for Mobileye don’t come along every quarter.
Of the second quarter’s roughly 40 AI M&A deals, I saw five that were squarely in the advertising and marketing space, with machine learning promising better customer engagement, sales enablement, and programmatic advertising.
A few other deals are worth noting because of cool technology, who the buyer was, who the seller was — or both. No financial terms were announced for any of the marketing-oriented AI deals.
Two of them focused on AI-powered data mining of social media — think of it as intelligent listening — with the ultimate goal of enhancing customer engagement.
The first was by a mobile and integrated payment solutions provider named Spindle, which acquired the AI assets of a book discovery website named CoverCake: i.e., data-mining algorithms that match content to users.
The second was by a customer care automation software company named Interactions that uses speech recognition and natural language translation. It acquired Digital Root, which “helps consumer brands qualify real-time content by using artificial intelligence to recognize, understand and prioritize opportunities and responses for sales, retention, and research,” according to the press release.
In other words, it figures out the best prospects for your sales team to target.
The buyer in the programmatic deal was a digital marketing services provider named Propel Media, which acquired DeepIntent Technologies. According to DeepIntent’s web site, its raison d’etre is to use AI to make programmatic marketplaces more contextually relevant: “Versed with the latest state-of-the-art algorithms for natural language modeling, DeepIntent is the world’s first media buying platform that can understand and target ads to fit the sentiment surrounding specific concepts (brands, people, places, things, and events) within publisher content.”
I’m always suspicious of “firsts.” Erik Matlick’s Industry Brains did a good job
matching text ads to content, adding a couple million dollars to my annual bottom line when I ran the TechWeb.com business at CMP Media circa 2001-2005. But the tech of course was much simpler, much
less precise, and the operating environment was narrower and better defined. DeepIntent’s AI presumably understands the actual meaning in any publisher’s content, rather than matching
keywords and taxonomical categories.
Next up: Sidetrade, a kind of ultra-modern British financial services consultancy that promises to use AI across-the-board, acquired C-Radar of France. C-Radar mines public databases, analyzes your own customer data, does profiling, etc., all with the ultimate goal of identifying and prioritizing your best sales leads.
The better-qualified the leads you focus on, the faster you get to generating cash. That’s the overall pitch behind all Sidetrade’s offerings and, I guess, why the C-Radar business makes for a sensible acquisition.
The fifth marketing-oriented AI deal was a merger between two small French digital agencies, NetBooster Group and Artefact, both of which boast a focus on AI. Their announcement promises AI “at the heart” everything they do: “It will be used to automate the creation of campaigns from strategy development through customer journey and from ad creation through campaign delivery and performance measurement."
The biggest price tag I could find on a second-quarter AI deal was $724 million. This was one of several deals that involved AI technology used to help airborne or land-bound drones/robots get around without bumping/crashing into things. The buyer was EriSat, a United Arab Emirates defense and aviation company that provides unmanned aerial solutions as well as manned.
The company acquired was Swiss-based Aerial Source which, according to the press release, “has the ability to autonomously take-off, land, and navigate while in-flight as well as recognize both surface and air targets, landmarks, and communicate with a network of joint UAS platforms during operations.”
Two other deals were notable because the buyer was Softbank, whose founder (you may have heard) has raised nearly $100 billion to invest in “moonshot” type visions, and the seller was Google (excuse me, Alphabet), which has been reevaluating how many moonshots a company can do and still keep its share price up around $1,000.
Softbank bought two AI/robotics companies that Google acquired in 2013 (when it was all still called Google) during a spree of AI/robotics deals that all seem to have lost their way when Android inventor Andy Rubin left to do his own thing.
Boston Dynamics is the famous one, mostly for making a demo video with eerily lifelike doggie robots that run around and keep their balance even when their human creators literally kick them around.
The second is humanoid robot maker SCHAFT, based in Japan, which spun out of a University of Tokyo lab before being acquired by Google.
You may have already read about Apple’s deal for Lattice Data, which uses an AI-based “inference engine” to make sense of unstructured data. Of course, everything has to be more dramatic in 2017, so some are using the term “dark data” these days instead of “unstructured” to enhance the dramatic effect.
But companies have used AI to make sense of unstructured text for literally decades. The first one I learned about was Autonomy’s clustering technology, about 25 or so years ago. Like all AI, this kind of tech has been advancing rapidly in the last few years. Lattice, founded in 2015, is a Stanford University spinout of a research project known as DeepDive. It counts among its founders a Stanford professor with a MacArthur Genius Grant, and one of the creators of Hadoop.
Several other AI deals to note: Microsoft, which acquired two small AI companies: one that automates cyber incident response (among about a half dozen AI cybersecurity deals) and the other that uses machine learning to help cloud computing users predict and manage their usage.
Spotify added one more AI-based content recommendation engine to the two similar technologies it acquired in the first quarter, and Airbnb acquired Trooly, which Airbnb has used to help it authenticate user identities since 2015. Trooly’s machine learning analyzes public databases and social media and, among other things, helps Airbnb figure out when you try to make side deals to avoid their fees.
That makes three dozen deals in the first quarter and 40 or more in the second, compared to only about 70 all last year.
I don’t believe AI is overhyped. It’s hyped, yeah. But not overhyped. How can anything possibly have as big an impact as AI proponents promise? People said the same thing about the internet in 1995.
Instead of hype, think of it as self-fulfilling prophecy. Nerds have vision; venture capitalists need to invest; companies are formed. Now, other companies that need to stay competitive are buying those companies, for real money. That’ll encourage more nerds and more VCs, while putting cash in the pockets of first-round nerds.
And so on, and so on, and so it goes.